Two-dimensional sinusoidal amplitude estimation with application to two-dimensional system identification (Q1851456)

The paper is devoted to amplitude estimation of two-dimensional sinusoidal signals from measurements corrupted by colored observation noise. The two-dimensional sinusoidal amplitude is investigated under the general frameworks of least-squares, weighted least-squares and matched-filterbank estimation. Various two-dimensional amplitudes are presented. The performances of these estimators for finite samples are compared numerically with one another as well as with the Cramer-Rao bound, the lower variance bound for any unbiased estimator. General guidelines for the selection and use of these estimators are provided. A new scheme for two-dimensional identification is introduced which has a closed-form expression. It is computationally simpler and statistically more accurate than the conventional output error method when the observation noise is colored. The Cramer-Rao bound for the two-dimensional system identification problem is also investigated in the paper. Some numerical examples are presented to illustrate the performances of the presented estimators.